3 ms·

> but the larger volume of TRISO fuel means the relative cost of dealing with it will be larger than their cost of dealing with spent LWR fuel Yes. But the mai

> but the larger volume of TRISO fuel means the relative cost of dealing with it will be larger than their cost of dealing with spent LWR fuel Yes. But the main cost of a reactor comes from the pressure vessel that the reactor is in. A LWR works at a pressure of about 160 bar (i.e. 160 times higher than the atmospheric pressure). The pressure inside a helium cooled reactor is about 50 bar. The cherry on the cake is that helium cooled reactors have much higher thermal efficiency (40% vs 30%). They can in principle even be used to produce hydrogen, in a much more efficient way than water electrolysis. In the US there is the Xe-100 design. But I doubt it will be commercial in less than 15 years. The Kairos design also uses TRISO pebbles, it's a molten salt design. That solves the pressure vessel problem even better. Still 15 years out probably.

pfdietz 2y ago
The pressure of a HTGR is lower, but because the temperature is higher, more expensive materials are needed. A LWR pressure vessel is within the creep limit of ordinary steel; HTGR outlet temperature is well above that limit (and I suspect in accident conditions the temperature goes even higher for passive dissipation of decay heat). This especially bites in applications proposing to use that high temperature industrially, such as in thermochemical water splitting. Also, I understand the passive safety of HTGRs is achieved by reducing the core thermal power density (and hence power density of decay heat). So for a given power, that core and pressure vessel will be much larger than in a LWR. If I'm reading a reference properly the ratio of power densities here is more than a factor of ten, which will more than balance the lower pressure and higher thermal efficiency. Material requirements for a pressure vessel scale as pressure x volume.
- credit_guy 2y ago
  > because the temperature is higher, more expensive materials are needed Is that a fact or an educated guess? Nuclear reactors do not use ordinary steel, they use nuclear grade steel. I don't know much about steel, but with a bit of googling I found out that the 316L steel (on of the 2 most common nuclear grade steels, the other being 304L) has much higher creep temperature than ordinary steel. This is the steel used in making the pressure vessels of PWRs and BWRs. The steel used in the HTR-10 (the precursor of the 2 HTR-PM rectors that China recently hooked to the grid) has a composition [1] that seems to me to be almost identical to the 316L [2]. The HTR-10 has an average helium temperature at the outlet of 700°C. Do you have some more concrete sources of information that would indicate the steel in HTGR is different and significantly more expensive than the steel in PWRs? [1] https://www-pub.iaea.org/MTCD/publications/PDF/te_1382_web/TE_1382_Part2.pdf https://www-pub.iaea.org/MTCD/publications/PDF/te_1382_web/T... [2] https://en.wikipedia.org/wiki/Marine_grade_stainless https://en.wikipedia.org/wiki/Marine_grade_stainless
  - pfdietz 2y ago
    Well, requiring a material to withstand a higher temperature is an added constraint, so of course it's going to increase cost. Otherwise, the more exotic material would have been used in the lower temperature application too. But that's all secondary to the observation you ignored, that the apples-to-apples comparison you're trying to make ignores power density, and when that's included HTGRs require much heavier pressure vessels.
    - credit_guy 2y ago
      > Well, requiring a material to withstand a higher temperature is an added constraint, so of course it's going to increase cost. This reasoning sounds good in theory, but practice and theory sometimes differ. The steel in reactors is subject to neutron embrittlement. In order to avoid that, you need special steel. Which happens to also withstand much higher temperatures. So no, it's not an additional constraint. > that the apples-to-apples comparison you're trying to make ignores power density. The strange thing is that the second part of your argument had nothing to do with TRISO. Not sure why you brought it up. You said that this is a tradeoff of passive safety. If you want passive safety, all things being equal, you need to accept lower power density. However, all things are not equal. When you use TRISO as a fuel, that can withstand absolutely huge temperatures, in excess of 2000°C. Since radiative cooling is governed by the Sefen-Boltzman equation (i.e. the radiated power is proportional to the 4th power of the temperature), if the pebbles reach high temperatures, they get rid of a lot of heat by radiation, which is then absorbed by the steel pressure vessel and conducted outside. Additionally, when you use TRISO in the form of pebbles, you recirculate the pebbles. You take them out and let them "rest", and then put them back later, so you limit the decay heat. Anyway, enough theory, let's look at actual numbers. The American designs Xe-100 and NuScale are perfect for our comparison. They both have a thermal output of 200 MW. NuScale is a PWR and has a lower efficiency, so the electrical output is 60 MW, while the Xe-100 produces 82.5 MWe. NuScale's reactor pressure vessel has a volume of 129 m3, and Xe-100 has 390 m3, so larger by a factor of 3. If we only care about the electricity output, then Xe-100 needs 4.73 m3/MWe, while NuScale 2.15 m3/MWe, the ratio is only 2.15. However, because the Xe-100 has to withstand much lower pressures, the mass of the pressure vessel is 274 tons, vs 260 tons for NuScale. Per MWe, the Xe-100 vessel is actually 23% lighter. [1] https://aris.iaea.org/Publications/SMR_Book_2020.pdf https://aris.iaea.org/Publications/SMR_Book_2020.pdf [2] https://aris.iaea.org/PDF/NuScale-NPM200_2020.pdf https://aris.iaea.org/PDF/NuScale-NPM200_2020.pdf
- petre 2y ago
  I guess we shall see how China's HTR enterprise goes. The Germans had a lot of problems with the AVR and the THTR. Probably part or why they ditched nuclear altogether. They also tend to overengineer things and add too much unecessary complexity, which is exactly what they did with the THTR. I'm also quite skeptic of carbon moderated reactors, the graphite blocks tend to crack, the pebbles create dust which can block the coolant flow and the graphite reflectors also crack and get contaminated with Cs/Sr. https://en.wikipedia.org/wiki/Pebble-bed_reactor https://en.wikipedia.org/wiki/Pebble-bed_reactor I saw the 4500 US$/kW cost figure for the HTR-PM. For the record, the AP1000 cost the Chinese around $2,000/kW in overnight capital costs (for the Sanmen and Haiyang plants). In the US the AP1000 construction costs are currently estimated around $6,800/kW and $4,500/kW for the following installed 10th unit. https://web.mit.edu/kshirvan/www/research/ANP193%20TR%20CANES.pdf https://web.mit.edu/kshirvan/www/research/ANP193%20TR%20CANE...

pfdietz 2y ago

The pressure of a HTGR is lower, but because the temperature is higher, more expensive materials are needed. A LWR pressure vessel is within the creep limit of ordinary steel; HTGR outlet temperature is well above that limit (and I suspect in accident conditions the temperature goes even higher for passive dissipation of decay heat). This especially bites in applications proposing to use that high temperature industrially, such as in thermochemical water splitting. Also, I understand the passive safety of HTGRs is achieved by reducing the core thermal power density (and hence power density of decay heat). So for a given power, that core and pressure vessel will be much larger than in a LWR. If I'm reading a reference properly the ratio of power densities here is more than a factor of ten, which will more than balance the lower pressure and higher thermal efficiency. Material requirements for a pressure vessel scale as pressure x volume.

credit_guy 2y ago
> because the temperature is higher, more expensive materials are needed Is that a fact or an educated guess? Nuclear reactors do not use ordinary steel, they use nuclear grade steel. I don't know much about steel, but with a bit of googling I found out that the 316L steel (on of the 2 most common nuclear grade steels, the other being 304L) has much higher creep temperature than ordinary steel. This is the steel used in making the pressure vessels of PWRs and BWRs. The steel used in the HTR-10 (the precursor of the 2 HTR-PM rectors that China recently hooked to the grid) has a composition [1] that seems to me to be almost identical to the 316L [2]. The HTR-10 has an average helium temperature at the outlet of 700°C. Do you have some more concrete sources of information that would indicate the steel in HTGR is different and significantly more expensive than the steel in PWRs? [1] https://www-pub.iaea.org/MTCD/publications/PDF/te_1382_web/TE_1382_Part2.pdf https://www-pub.iaea.org/MTCD/publications/PDF/te_1382_web/T... [2] https://en.wikipedia.org/wiki/Marine_grade_stainless https://en.wikipedia.org/wiki/Marine_grade_stainless
- pfdietz 2y ago
  Well, requiring a material to withstand a higher temperature is an added constraint, so of course it's going to increase cost. Otherwise, the more exotic material would have been used in the lower temperature application too. But that's all secondary to the observation you ignored, that the apples-to-apples comparison you're trying to make ignores power density, and when that's included HTGRs require much heavier pressure vessels.
  - credit_guy 2y ago
    > Well, requiring a material to withstand a higher temperature is an added constraint, so of course it's going to increase cost. This reasoning sounds good in theory, but practice and theory sometimes differ. The steel in reactors is subject to neutron embrittlement. In order to avoid that, you need special steel. Which happens to also withstand much higher temperatures. So no, it's not an additional constraint. > that the apples-to-apples comparison you're trying to make ignores power density. The strange thing is that the second part of your argument had nothing to do with TRISO. Not sure why you brought it up. You said that this is a tradeoff of passive safety. If you want passive safety, all things being equal, you need to accept lower power density. However, all things are not equal. When you use TRISO as a fuel, that can withstand absolutely huge temperatures, in excess of 2000°C. Since radiative cooling is governed by the Sefen-Boltzman equation (i.e. the radiated power is proportional to the 4th power of the temperature), if the pebbles reach high temperatures, they get rid of a lot of heat by radiation, which is then absorbed by the steel pressure vessel and conducted outside. Additionally, when you use TRISO in the form of pebbles, you recirculate the pebbles. You take them out and let them "rest", and then put them back later, so you limit the decay heat. Anyway, enough theory, let's look at actual numbers. The American designs Xe-100 and NuScale are perfect for our comparison. They both have a thermal output of 200 MW. NuScale is a PWR and has a lower efficiency, so the electrical output is 60 MW, while the Xe-100 produces 82.5 MWe. NuScale's reactor pressure vessel has a volume of 129 m3, and Xe-100 has 390 m3, so larger by a factor of 3. If we only care about the electricity output, then Xe-100 needs 4.73 m3/MWe, while NuScale 2.15 m3/MWe, the ratio is only 2.15. However, because the Xe-100 has to withstand much lower pressures, the mass of the pressure vessel is 274 tons, vs 260 tons for NuScale. Per MWe, the Xe-100 vessel is actually 23% lighter. [1] https://aris.iaea.org/Publications/SMR_Book_2020.pdf https://aris.iaea.org/Publications/SMR_Book_2020.pdf [2] https://aris.iaea.org/PDF/NuScale-NPM200_2020.pdf https://aris.iaea.org/PDF/NuScale-NPM200_2020.pdf
petre 2y ago
I guess we shall see how China's HTR enterprise goes. The Germans had a lot of problems with the AVR and the THTR. Probably part or why they ditched nuclear altogether. They also tend to overengineer things and add too much unecessary complexity, which is exactly what they did with the THTR. I'm also quite skeptic of carbon moderated reactors, the graphite blocks tend to crack, the pebbles create dust which can block the coolant flow and the graphite reflectors also crack and get contaminated with Cs/Sr. https://en.wikipedia.org/wiki/Pebble-bed_reactor https://en.wikipedia.org/wiki/Pebble-bed_reactor I saw the 4500 US$/kW cost figure for the HTR-PM. For the record, the AP1000 cost the Chinese around $2,000/kW in overnight capital costs (for the Sanmen and Haiyang plants). In the US the AP1000 construction costs are currently estimated around $6,800/kW and $4,500/kW for the following installed 10th unit. https://web.mit.edu/kshirvan/www/research/ANP193%20TR%20CANES.pdf https://web.mit.edu/kshirvan/www/research/ANP193%20TR%20CANE...